Party Discipline and Government Spending: Theory and Evidence

Transcription

1 Party Discipline and Government Spending: Theory and Evidence Marta Curto-Grau Galina Zudenkova Preliminary draft. Please do not cite. Abstract Political parties value highly the commitment of their members with their party line as internal cohesion helps them build reputation. Excessive party discipline at the individual level is, however, electorally costly and, thus, in some instances legislators may be reluctant to vote in agreement with their party. In this paper we examine theoretically and empirically the possibility that party leaders systematically use distributive spending to reward discipline within their party. The empirical analysis is based on U.S. data on the geographical distribution of discretionary spending from the federal government to congressional districts and the level of party loyalty of legislators in the House of Representatives. Using a panel data analysis with time and district fixedeffects, we find that an increase in loyalty in congressional roll-call voting translates into larger discretionary spending. However, this effect is only statistically significant in districts represented by members in the House majority. Keywords: Party discipline; Majority party; Minority party; Party leadership; Porkbarrel spending. JEL classification: D72, H41 We are grateful to Lutz Brückner for research assistance. We also thank Ken Sheplse and James Snyder for their comments and advice. An earlier - and very different - version of this paper circulated with the title Party Discipline in Congress. Research Center for Distributional Conflict and Globalization, University of Heidelberg, Bergheimer Str. 58, Heidelberg, marta.curto@awi.uni-heidelberg.de Department of Economics, University of Mannheim, D Mannheim, Germany. address: galina.zudenkova@gmail.com. University of Mannheim 1

2 1 Introduction In modern democracies, party discipline is defined as the ability of the leadership of a party to control its members in the legislature. Such control is often difficult to achieve because in the effort to serve their constituents, individual legislators may try to defect their party line. Excessive loyalty with the party may have large negative consequences at the polls. For this reason, to soften electoral punishment and foster discipline, party leaders may reward loyalty. Such rewards may include favorable committee assignments and leadership positions, campaign funds, district visits by party notables, federal projects targeted to a member s district, expedited treatment for a member s favorite bills, and invitations to serve as speaker pro tem (Snyder and Groseclose (2000), p.194). Applied studies in the political science literature have emphasized the political consequences of party discipline. Some of them examine the electoral punishment that too loyal legislators suffer at the polls (e.g., Carson and Young (2010)). Others shown that party influence affects legislators voting behavior (Heller and Mershon (2008); Krehbiel (2000); McCarty and Rosenthal (2001); Rohde (1991); Snyder and Groseclose (2000)). For the U.S, Barber and Godbout (2014) show that majority party leaders reward loyalty in roll call voting through larger campaign contributions for the next election. Further, a number of authors have produced formal models of party discipline. Some researchers have elaborated on informational arguments, pointing out that strong party discipline informs voters about the future policy of a candidate who, once elected, cannot deviate from the partys official platform (Ashworth and Bueno de Mesquita (2004); Castanheira and Crutzen (2010); Cox and McCubbins (1993); Snyder and Ting (2002)). In a similar vein, Grossman and Helpman (2008) defined party discipline as a party s ability to induce ex-post adherence to a pre-announced position. In other words, the level of party discipline determines the "extent of commitment to party platforms." 1 In some other studies, party discipline has been modeled as the ability of the party leadership to control its members in the legislature such that they vote in line with the party s ideological position (Colomer (2005); Eguia (2011); Iaryczower (2008); McGillivray (1997); Patty (2008); Volden and Bergman (2006)). In these models, the party leaders objective is to discipline party members who might have different ideological preferences. Diermeier and Feddersen (1998a,b) provided an institutional explanation for cohesive voting of legislators in parliamentary systems. They showed that the vote-of-condence procedure common in parliamentary democracies creates an incentive for cohesion in voting. In any case, party discipline is a valuable asset to party leaders because 1 Grossman and Helpman (2008), p

3 it acts as a signal of cohesion which, in turn, helps building a political brand (Cox and McCubbins (1993). While a number of studies have identified and formally analyzed party discipline, there remains a great deal of uncertainty as to how discipline affects actual policy outcomes 2. The present paper contributes to this line of research. In particular, we seek to explain theoretically and empirically the effects of party discipline on the geographic allocation of discretionary spending 3, an issue largely disregarded by existing studies on distributive politics. The relationship between party discipline and discretionary spending is not a priori clearcut. Primo and Snyder (2010) suggest that party strength may be a possible solution to the problem of inefficient universalism (p. 354), thus predicting lower federal spending under a strong party system. Their main argument is that when parties have strong electoral organizations [i.e. party strenght outside the legislature], legislators have less incentives to engage in credit-claiming activities, such as constituency service and particularistic spending, because the party organization will do much of the work in helping them run for office. Although party discipline may reduce wasteful spending, Ashworth and Bueno de Mesquita (2006) have noted a fact that is at odds with this hypothesis: since the early 1990s, the U.S. has experienced an increase both in party cohesion in congressional roll-call voting and porkbarrel spending (measured as congressional earmarks in appropriation bills). This pattern is more consistent with the idea of pork as a reward for loyalty. Pearson (2008) suggests that this may be indeed the case: When party leaders in both chambers finalize the details of major legislation, they have opportunities to reward loyalty. The increasing number of earmarks added in conference committee or during budget summit negotiation may provide limited opportunities to reward loyalty that are worth further investigation (p. 113) 4. The U.S. offers a good setting to test the hypothesis that party leaders favor loyal members 2 An exception is Grossman and Helpman (2008) who studied how differences in party discipline affect local public goods spending. In their setting, however, party discipline is modeled as an exogenously given institutional variable -an "extent of commitment to party platforms." They found a non-monotonic relationship between this institutional variable and pork-barrel spending. The present paper in turn adopts a standard denition of party discipline as the ability of party leaders to induce members to support the party line. Moreover, it endogenizes party discipline by modeling explicitly the party leaders problem. 3 The term discretionary spending refers to non programmatic spending that has a partisan bias and is targeted to certain constituencies. Henceforth we use interchangeably the terms discretionary spending and pork barrel -an equivalent term commonly used in studies on the U.S. 4 Moreover, there is anecdotal evidence in the media as, for instance, the claim of a journalist writing about the need to eliminate a ban on earmarks adopted in 2010: "It s time to bring back the earmark, and to treat it for what it is: a reward for party loyalty. Congress-Do-Something 3

4 through large amounts of federal spending. In this country party discipline is relatively lower than in other countries and shows considerable variation over time. To conduct our analysis, we use data on discretionary spending from the federal government to congressional districts and party discipline in the U.S. House of Representatives over the period Regarding our measure of party discipline, we focus our analysis on party loyalty in policy voting, that is, the behavior of parties (and their members) within the legislative process (in contrast to Primo and Snyder (2010), probably the closest paper to ours). We measure this type of discipline using the party unity scores published yearly by Congressional Quarterly. These scores are based on roll call votes in which a majority of Democrats opposed a majority of Republicans; unity scores are measured as the percentage of such votes in which a member of the Congress voted in line with her party. The estimated effects are obtained using ordinary least squares with time and district fixed effects in order to reduce the potential bias generated by omitted variables. The results indicate that a 10 unit increase in the level of party discipline increases the federal spending a congressional district receives by approximately 6% (for an average district in our sample this represents $80 per capita). Most importantly, this effect is only statistically significant for districts represented by a member of the House majority, which points to an exclusive advantage of the majority party to reward the loyalty of their members. In our period under analysis, the House of Representatives has been under the control of both Republicans and Democrats, ruling out the possibility that such advantage is inherent in only one party. Lastly, we also test for differential effects in marginal versus safe districts but we do not observe a statistically significant pattern. The rest of the paper is organized as follows. Section 2 outlines the institutional background of the paper (i.e. the role of party leaders and the evolution of party discipline in the U.S. Congress) and a set of theoretical considerations that are relevant for our analysis. Section 3, presents the empirical analysis and discusses our findings. Finally, Section 4 concludes. 2 Institutional and theoretical considerations Party discipline and the role of party leaders in the U.S. Congress In the U.S., party organizational strength experienced a sharp decline by the mid-1970s (Mayhew (1986)). This fact took place in parallel with an increase in legislative loyalty in the Congress (i.e. it became more common for legislators to vote with their party line). As previous literature has noted, party discipline is a difficult concept to measure quantita- 4

5 tively (see e.g. Snyder and Groseclose (2000); Krehbiel (2000)). However, it seems reasonable to believe that party discipline becomes especially visible when members of the Congress have to vote on an issue in which a majority of one party votes against a majority of the other party - a concept known as party unity votes 5. A common measure of the level of party discipline of an individual legislator on a given year is, thus, the percentage of party unity votes on which a member voted yea or nay in agreement with his or her party. As shown in Figure 1, this score was on average around 60% by the early 70s but it raised significantly during the following decades and nowadays it amounts to 80%-90%. 5 Snyder and Groseclose (2000) estimate the extent to which party influences roll-call votes in the U.S. Congress for the period and argue that lopsided roll calls do not need to be better predictors of party influence. However, they also admit that this finding may be a statistical artifact. 5

6 Figure 1: Party discipline and party control in the U.S. Congress Average Unity Scores Panel A: U.S. House of Representatives Fiscal Year Average Unity Scores Panel B: U.S. Senate Fiscal Year Democratic Majority Democratic Score Republican Majority Republican Score Source: Congressional Quarterly, Annual Report, January 2011 In the U.S., party leaders play an important role within the legislative process. Once elected, they can assist (and favor) their party members in several ways. Party leaders, however, have not always had such a relevant role. In fact, it was not until the end of the 19th century that the congressmen of each political party organized and assembled separately to elect some members who would represent them, and until the 20th century the posts of 6

7 majority and minority leaders were not official political positions 6. In the 1970s the House of Representatives undertook a series of reforms that partitioned the postwar period into a Weak Party Era (prior to the reforms) and a Strong Party Era (from the mid-70s until nowadays). Hall and Shepsle (2013) examine how this changes affected the value of seniority in the U.S. House. Their study is based on the fact that the rule changes undertaken by the Democratic and Republican party during the 70s increased the power of the party leadership at the expense of senior members. The findings of their study suggest that the electoral rewards to seniority are higher during the Weak Party Era, when senior members are able to convert their longevity into more value for their constituents. The reason behind this finding is that during this era the power was in the hands of committee chairs (who obtained such positions through seniority) but after the 70s the power shifted towards party leaders. As previously discussed, nowadays party leaders aim at reaching the highest possible degree of unified party action and to achieve this goal they may reward loyal members during the legislative process. In addition, they may want to compensate for the severe electoral consequences that party loyalty has for incumbents running for re-election. An empirical analysis for the U.S. shows that voters punish elected representatives not for their ideological extremism but for voting too often with their party on salient issues that divide Republicans and Democrats (Carson and Young (2010)). While Carson and Young (2010) suggest that the reason why legislators continue supporting their party line most of the time -despite its electoral consequences- is a combination of personal conviction and support for the collective interests of their parties, it seems also plausible to think that loyal legislators receive monetary benefits (e.g. campaign contributions or pork-barrel spending) in exchange for discipline. Furthermore, Pearson (2005) shows that there are other legislative resources to benefit loyal party members. More specifically, she analyzes how party discipline affects which members bills, amendments, resolutions, and suspensions were considered for the floor debate in the House of Representatives during the years Theoretical arguments and hypothesis Extant studies on the political determinants of pork-barrel spending can be divided broadly in three main groups. One of them emphasizes the bias in the allocation of funds towards districts represented by senior representatives (e.g.debacker (2011); Bernhardt (2004)) 7. Another set of studies examine whether government spending is targeted to swing or core The intuition behind these papers is that, over the years in office, senior legislators have established personal relationships that help them attract funds. 7

8 constituencies 8. A third group focuses on how congressional majorities or the alignment with the President s party may determine the distribution of public spending (e.g., Levitt and Snyder (1995); Berry (2010); Dynes and Huber (2015)). A fourth group analyzes whether politically aligned districts are disproportionately favored by upper-level governments in the distribution of discretionary grants (e.g. Arulampalam and Dutta (2009); Brollo and Nannicini (2012)). In any case, as discussed in the Introduction, very little is known about the potential effect of party discipline on the geographic allocation of federal funds. We believe that the paper that is closest to ours is Primo and Snyder (2010), which examines the relationship between party strength and federal spending in the U.S. Their study pays special attention to party organizational strength (i.e. strength outside the legislature), but they also examine briefly the influence of legislative party strength (i.e. discipline inside the legislature) on the distribution of federal spending. We depart from their study in several ways. First, in their analysis of the role of legislative party strength, they limit their time frame to the 1950s and 1960s given that they rely on a measure of party discipline only available for that time frame. As explained before, the role of party leaders in the Congress has changed greatly after the 1970s and, thus, in our setting party leaders have the capacity to reward loyalty. Second, as a measure of legislative discipline they use an index constructed by Zeller (1954), who using a measure of party cohesion in the legislature classified U.S. states into states with strong or weak parties. Therefore, party discipline is measured as a binary variable that does not change over time within states. In contrast, to this approach we use a measure (party unity scores) that varies over time for each legislator. Third, they focus their analysis on total federal spending to states, while we examine the distribution of discretionary spending that can be more easily targeted to a particular district and is more susceptible to manipulation. Regardless of whether party leaders use discretionary spending to reward loyalty or they use it to offset electoral defeats, our hypothesis is that, holding other things equal, an increase in the level of party discipline should translate into an increase in pork-barrel spending to a district. Another feature that is noteworthy in this context is the distinction between the majority and minority parties. Previous research (e.g. Albouy (2013); Dynes and Huber (2015); Levitt and Snyder (1995)) finds that members of the majority party in the U.S. House garner additional federal spending in their districts suggesting that the majority party has an advantage in terms of proposal, or agenda-setting power. Therefore, although both the 8 See, e.g.,lindbeck and Weibull (1987); Dixit and Londregan (1996); Cox and McCubbins (1986). For a summary of the mixed evidence found by this strand of literature see Golden and Min (2013). 8

9 majority and minority leaders may value equally the loyalty of their party members, it is likely that the minority leader lacks the legislative resources to reward discipline. Moreover, it is likely that the majority leader is reluctant to send large amounts of pork to opposition districts given that without party alignment it may be difficult to claim credit for that. All in all, our hypothesis is that the returns to party discipline should be considerably larger for majority party members. Despite the above hypothesis, it should be noted that the presence of such bias does not prevent minority states from attracting pork-barrel spending provided that there exists a partisan blame avoidance. In other words, to inoculate itself against minority party criticism, the majority concedes a share of distributive projects to the minority and thereby renders partisan distinctions inconsequential in the awarding of earmarks. (Balla and Sigelman (2002), p.518) 3 Empirical application In this Section, we provide empirical evidence on the relationship between party discipline in the U.S. House of Representatives and the geographic distribution of pork-barrel spending among congressional districts over the years Provided that representatives in the U.S. Congress show a relatively large degree of autonomy, we believe that this country provides a good setting for testing the hypothesis that party discipline is rewarded through distributive spending. Moreover, the levels of loyalty within the legislature have changed remarkably over time within congressional districts. This allows us to use a panel data analysis with district fixed-effects. 3.1 Data Pork-barrel spending To focus our analysis on the distribution of pork-barrel spending we follow previous work on the U.S. that classifies spending programs into high variance and low variance programs based on the coefficient of variation of each program (DeBacker (2011); Levitt and Snyder (1995); Dynes and Huber (2015)). Information on federal spending is obtained from the Federal Assistance Award Data System (FAADS), which has data on federal transfers disaggregated at the congressional district level. Although FAADS data represent a smaller percentage of the federal budget than data from the Consolidated Federal Funds Report -an alternative source-, the former is preferred 9

10 as it is more disaggregated and identifies with more precision the recipient of the funds. The coefficient of variation used to capture discretionary spending is the variance in mean awards by each program across congressional districts. High-variance spending is commonly associated with discretionary spending as it is usually not formula-based and it tends to fund smaller programs that are more susceptible of targeting. In our paper, we use the same FAADS data than Dynes and Huber (2015) 9 and, thus, we take the same rule of thumb to select high-variance spending 10. This is arguably not a perfect measure of pork-barrel spending, however, pure distributive spending (i.e. earmarks in appropriation bills) is only available at the state level and an aggregation of party discipline at such level would dismiss important information given the large divergences within states. It is important to note that the federal budget of a certain fiscal year (FY) is approved during the previous year. This means, for instance, that pork-barrel expenditures of FY1993 were passed in 1992 by legislators elected in the 1990 election. In our sample, congressional districts receive on average $1151 per capita of discretionary spending. Party discipline With regards to our main explanatory variable of interest, in our analysis we use a measure of party loyalty in policy voting. This measure is the party unity score of each legislators in a given year (i.e. the percentage of party unity votes on which a member voted yea or nay in agreement with his or her party). Data on party unity scores are collected from Congressional Quarterly Allmanac. In our sample, this is a score has a minimum of 9.36 and a mean of Table1 below summarizes the descriptive statistics for the variables described in this section. Additional control variables Besides party discipline, our explanatory variables also include additional political and economic factors that vary over time within a congressional district. The political controls are based on previous literature on distributive spending and include a large set of variables. Three of these variables help identifying whether a legislator holds a certain leadership position that would help her attracting larger funds: Committee Chair, Member of Appropriations 9 Replication files are kindly made available online in 10 The threshold they use to split spending into high and low variation is 1 (which is a natural break at the lower end of a histogram of the coefficients of variation). Therefore, programs with a coefficient of variation (CV) above this threshold are considered to be pork-barrel spending. The online appendix of their paper contains detailed information on how the data is collected and how the CV is computed. 10

11 Committee, Member of Ways and Means Committee, and Member of Majority-Party leadership are binary variables that equal to one if the representative is a chair of a committee, a member of the Appropriations committee, a member of the Ways and Means committee, or a member of the Majority-Party leadership, respectively. Also, we add two variables to account for the possibility that members with a certain partisanship have an agenda-setting advantage that helps secure funding: this would be the case of members belonging to the President s party (Member of President s Party) or members of the House majority (Member of House Majority), which does not necessarily coincide with the President s party. In addition, to account for possible taste effects, we include in our regressions a binary variable that equals one if a member is republican (Republican) 11. The years in office are also an important determinant of the distribution of funds as more senior members are meant to be in a better position to attract grants; thus, we include a dummy variable that indicates wether a representative is in his first term (Freshman), and a variable measuring the number of years an incumbent has been in office (Seniority). The literature on pork-barrel politics also suggests that electoral competition in a district influences the allocation of government spending. In line with this argument, we add a continuous variable for the district s 2 party vote margin for the winning presidential candidate (Winning Presidential Candidate s Margin) and a dummy indicating whether a representative s last election margin was less than 5% (Close Election). Finally, we include a binary variable equal to one if a representative is a ranking minority member of a committee (Ranking Minority). Even if the spending we are analyzing is to a large extent discretionary, economic variables may also influence its geographical allocation across districts. Hence, our econometric specifications include a list of socioeconomic variables that capture changes in spending needs. Such variables are unemployment rate, population density, population size, mean household income, the percentage of population that is black, sixty five or over. These variables are obtained from the Decennial Census 12, which is the only source with data at the congressional district level. As the Decennial Census is not published yearly we use the censuses from 1980, 1990, 2000 and 2010 and interpolate the data to obtain yearly observations. 11 Republican legislators tend to secure less grants for their districts

12 Table 1: Descriptive statistics Variable Mean St.Dev. Min Max High-Variation Spending per capita Party unity Member of President s Party Committee chair Member of Majority-Party leadership Member of Appropriations Committee Member of Ways & Means Committee Republican Freshman Close election (=1 if vote margin < 5%) Winning presidential candidate s margin in district Member of House Majority Ranking minority member on committee Population Black population Unemployment rate Population density Total population Econometric model We estimate the relationship between party discipline and pork-barrel spending using a panel data model with district and time fixed-effects as expressed in the following equation: y it = βp arty Unity it + γp olitics it + θsocioecon it + µ i + δ t + u it (1) where y it is High-Variance spending per capita in district i at fiscal year t, in 2010 dollars, adjusted using consumer price index; P olitics it and Socioecon it are the set of political and socioeconomic controls, respectively (see previous subsection for details); δ t and µ i represent year and district effects, respectively; u it is an error term (which is clustered at the congressional district level to provide consistent estimates); P artyunity it is the level of party discipline. According to our hipothesis, β is expected to be positive given that party discipline is valued positively by party leaders who afterwards reward it. 12

13 3.3 Results Table 2 below shows the OLS estimates obtained from regressing Equation (1) with several modifications: the regression in column (2) includes an interaction between the level of party discipline and a binary variable for close elections; column (3) includes a linear time trend to test whether the effect of party discipline varies along the years or remains constant for the whole period under analysis. Despite the lack of statistical significance of our main explanatory variable of interest, the results in Table 2 show that there are two political variables that are important determinants of the distribution of discretionary spending across districts: whether a legislator s last election was close and whether the legislator is a member of the majority party. In both cases, the coefficients are positive, in line with the findings of previous studies. 13

14 Table 2: Party discipline and discretionary spending to U.S. congressional districts, (1) (2) (3) Party unity (1.384) (1.431) (1.847) Party unity x Close election (2.849) Party unity x Trend (0.169) Close election (37.790)*** ( ) (37.835)*** President s party (47.587) (47.583) (47.560) Majority party (29.244)*** (29.353)*** (29.682)*** Margin at presidential elections (1.702) (1.702) (1.712) Majority leadership ( ) ( ) ( ) Committee chair (59.987) (59.946) (60.076) Ranking minority member on committee (68.568) (68.556) (68.647) Appropriations Committee (65.088) (65.310) (65.257) Ways & Means Committee (95.989) (95.999) (96.113) Republican (54.517) (54.688) (54.802) First term (27.941) (27.829) (27.913) Constant 11, , , (2, )*** (2, )*** (2, )*** Observations 10,359 10,359 10,359 R-squared Mean of outcome in sample 1,290 1,290 1,290 # Unique districts 1,564 1,564 1,564 Notes: Robust standard errors in parentheses, ***p<0.01, ** p<0.05, * p<0.1. SE clustered by congressional district. Year dummies in all equations. Dependent variable =High-Variance Spending per capita. Additional socioeconomic controls included are population, unemployment rate, percentage of population over 65, percentage of black population, median household income, and population density. 14

15 Given that the minority party may not have access to resources to reward party discipline, as explained in a previous section, we estimate Equation (1) for two different sub-samples: districts represented by a legislator in the majority party (Panel A) and districts represented by minority legislators (Panel B). The results in Column 1 show that the effect of party discipline on government spending is statistically significant only when the representative is in the House majority. In this case, a 10 unit increase in the level of party discipline increases the federal spending a congressional district receives by 80 dollars per capita. This represents a 6% increase for the average distict. Columns (2) and (4) replace the variable Party unity for a binary variable (Disciplined member) that equals to one if the level of Party Unity of a legislator if above the average. In this case, according to the results in column (2), disciplined members receive on average 11% more pork per capita than non-disciplined members. 15

16 Table 3: Party discipline and discretionary spending. Majority versus minority districts. (1) (2) (3) (4) Panel A. Majority party districts Panel B. Minority party districts Party unity (4.008)** (2.040) Disciplined member (71.661)** (36.390) Close election (75.092)*** (74.395)*** (46.888)*** (46.646)*** President s party ( ) ( ) ( ) ( ) Margin at presidential elections (2.617) (2.655) (2.942) (2.934) Majority leadership (52.600)*** (49.695)*** ( ) ( ) Committee chair (69.494) (69.569) ( ) ( ) Appropriations Committee ( ) ( ) (67.217) (66.550) Ways & Means Committee ( ) ( ) ( ) ( ) Republican ( )* ( )* ( ) ( ) First term (46.873) (46.846) (36.870) (36.698) Constant 9, , , , (2, )*** (2, )*** (5, )** (5, )** Observations 5,791 5,791 4,568 4,568 R-squared # Unique districts 1,297 1,297 1,093 1,093 Notes: Robust standard errors in parentheses, ***p<0.01, ** p<0.05, * p<0.1. SE clustered by congressional district. Year dummies in all equations. Dependent variable =High-Variance Spending per capita. Additional socioeconomic controls included are population, unemployment rate, percentage of population over 65, percentage of black population, median household income, and population density. Lastly, in Table 4 we test the relationship between party discipline and non-discretionary spending. In this case the dependent variable is low-variance spending and, as expected, the coefficients of Party Unity and Disciplined Member are not statistically significant given that this type of spending is much less susceptible to political manipulation. 16

17 Table 4: Party discipline and non-discretionary spending to U.S. congressional districts, (1) (2) (3) (4) Panel A. Majority party districts Panel B. Minority party districts Party unity (1.516) (3.974) Disciplined member (23.100) (67.269) Close election (30.073) (29.954) (30.629)** (30.210)** President s party -1, , , , (65.554)*** (65.589)*** ( )*** ( )*** Margin at presidential elections (1.278) (1.280) (1.307) (1.306) Majority leadership (24.010) (22.656) (57.281) (58.318) Committee chair (35.922) (35.689) (59.204) (58.051) Appropriations Committee (36.695) (36.964) (50.110) (46.342) Ways & Means Committee (69.459) (69.644) (45.317) (44.411) Republican -1, , , , (64.231)*** (64.226)*** ( )*** ( )*** First term (21.369) (21.388) (38.744)* (39.281)* Constant 31, , , , (1, )*** (1, )*** (1, )*** (1, )*** Observations 5,791 5,791 4,568 4,568 R-squared # Unique districts 1,297 1,297 1,093 1,093 Notes: Robust standard errors in parentheses, ***p<0.01, ** p<0.05, * p<0.1. SE clustered by congressional district. Year dummies in all equations. Dependent variable=low-variance Spending per capita. Additional socioeconomic controls included are population, unemployment rate, percentage of population over 65, percentage of black population, median household income, and population density. 17

18 4 Conclusions This paper analyzes the relationship between party discipline in the Congress and the geographic allocation of pork-barrel spending across districts. The role of party discipline in the allocation of pork has been largely omitted in the existing literature on distributive politics. However, there are reasons to believe that party leaders may wish to reward loyalty through targeted spending. Strong party discipline within the legislature is highly valued by party leaders given that it increases the reputation of the party. Loyalty has been proven to be rewarded with campaign contributions (Barber and Godbout, 2014) and consideration of bills for the floor debate in the House of Representatives (Pearson, 2005). Moreover, excessive party loyalty has been found to have negative electoral consequences and to offset such situation party leaders may compensate legislators with monetary benefits (such as pork-barrel for their districts). Our hypothesis is tested empirically using congressional district level data on federal spending and party discipline in policy voting in the U.S. House of Representatives over the period The U.S. is especially suitable for examining this issue given that its congressmen exhibit a relatively large degree of autonomy. Moreover, the levels of party discipline have varied considerably from the 1970s, when party organizational strength (i.e party discipline outside the legislature) collapsed but party discipline within the legislature started to increase sharply. As for the measure of party discipline used in this study, we rely on the so-called party unity scores, i.e. the percentage of votes in which a member of Congress votes in agreement with his or her party on roll call votes where the majority of one party opposes the majority of the other. The results using a panel data analysis with district fixed-effects reveal a positive relationship between party discipline and discretionary spending per capita. These effects are only statistically significant for districts represented by members in the House majority, whose party leader is likely to enjoy a relative legislative advantage to distribute expenditures. Moreover, no effect is found when analysing non-discretionary spending. References Albouy, David Partisan Representation in Congress and the Geographic Distribution of Federal Funds. Review of Economics and Statistics 95(1): Arulampalam, W., S. Dasgupta A. Dhillon and B. Dutta Electoral Goals and Center- 18

19 State Transfers: a Theoretical Model and Empirical Evidence from India. Journal of Development Economics 88: Ashworth, Scott and Ethan Bueno de Mesquita Endogenous Party Discipline with Variable Electoral and Legislative Institutions. Institute of Governmental Studies Working Paper Ashworth, Scott and Ethan Bueno de Mesquita Delivering the Goods: Legislative Particularism in Different Electoral and Institutional Settings. The Journal of Politics 68(1): Balla, Steven J., Lawrence Eric. D. Maltzman F. and Lee Sigelman Partisanship, Blame Avoidance, and the Distribution of Legislative Pork. American Journal of Political Science 46(3): Barber, M., Canes-Wrone Brandice and Jean-François Godbout Party Loyalty and Campaign Contributions. In Paper presented at the European Political Science Association Annual Meeting, 21 June Bernhardt, Dan, Dubey-Sangita Hughson Eric Term Limits and Pork Barrel politics. Journal of Public Economics 88 (12): Berry, Christopher R., Burden Barry C. Howell William G The President and the Distribution of Federal Spending. American Political Science Review 104(4). Brollo, F. and T. Nannicini Tying your Enemy s Hands in Close Races: the Politics of Federal Transfers in Brazil. American Political Science Review 106: Carson, Jamie L., Koger Gregory. Lebo Matthew J. and Everett Young The Electoral Costs of Party Loyalty in Congress. American Journal of Political Science 54(3): Castanheira, Micael and Benoit S. Y. Crutzen Comparative Politics with Endogenous Intra-Party Discipline. mimeo. Colomer, Josep M Policy Making in Divided Government: A Pivotal Actors Model with Party Discipline. Public Choice 125: Cox, Gary W and Mathew D. McCubbins Legislative Leviathan: Party Government in the House. Berkeley: University of California Press. 19

20 Cox, G.W. and M McCubbins Electoral Politics as a Redistributive Game. Journal of Politics 48(2): DeBacker, Jason The Price of Pork: The Seniority Trap in the U.S. House. Journal of Public Economics 95: Diermeier, Daniel and Timothy J. Feddersen. 1998a. Cohesion in Legislatures and the Vote of Confidence Procedure. American Political Science Review 92: Diermeier, Daniel and Timothy J. Feddersen. 1998b. Comparing Constitutions: Cohesion and Distribution in Legislatures. European Economic Review 42: Dixit, A and J. Londregan The determinants of success of special interests in redistributive politics. Journal of Politics 58: Dynes, Adam M. and Gregory A. Huber Partisanship and the Allocation of Federal Spending: Do Same-Party Legislators or Voters Benefit from Shared Party Affiliation with the President and the House Major. American Political Science Review 109(1). Eguia, Jon X Voting Blocs, Party Discipline and Party Formation. Games and Economic Behavior 73: Golden, M. and B. Min Distributive Politics around the World. Annual Review of Political Science 16: Grossman, Gene M. and Elhanan Helpman Institutions and Economic Performance. Cambridge: Harvard University Press chapter Party Discipline and Pork-Barrel Politics. Hall, Andrew B. C and Kenneth A. Shepsle The Changing Value of Seniority in the U.S. House: Conditional Party Government Revised. The Journal of Politics pp Heller, William B. and Carol Mershon Dealing in Discipline: Party Switching and Legislative Voting in the Italian Chamber of Deputies, American Journal of Political Science 52: Iaryczower, Matias Contestable Leadership: Party Leaders as Principals and Agents. Quarterly Journal of Political Science 3: Krehbiel, Keith Party Discipline and Measures of Partisanship. American Journal of Political Science 44:

21 Levitt, Steven D. and James M. Snyder Political Parties and the Distribution of Federal Outlays. American Journal of Political 39 (4): Lindbeck, A. and J. W. Weibull Balanced-Budget Redistribution as the Outcome of Political Competition. Public Choice 52: Mayhew, David Placing Parties in American Politics. Princeton, NJ: Princeton University Press. McCarty, Nolan, Keith T. Poole and Howard Rosenthal The Hunt for Party Discipline in Congress. American Political Science Review 95: McGillivray, Fiona Party Discipline as a Determinant of the Endogenous Formation of Tariffs. American Journal of Political Science 41: Patty, John W Equilibrium Party Government. American Journal of Political Science 52: Pearson, Kathryn Party Discipline in the Contemporary Congress:Rewarding Loyalty in Theory and in Practice PhD thesis University of California, Berkeley. Pearson, Kathryn Why Not Parties?: Party Effects in the United States Senate. Chicago: University of Chicago Press chapter Party Loyalty and Discipline in the Individualistic Senate, pp Primo, David M. and James M. Snyder Party Strength, the Personal Vote, and Government Spending. American Journal of Political Science 54(2): Rohde, David Parties and Leaders in the Postreform House. Chicago: University of Chicago Press. Snyder, James M. Jr. and Michael M. Ting An Informational Rationale for Political Parties. American Journal of Political Science 46: Snyder, James M. Jr. and Tim Groseclose Estimating Party Influence in Congressional Roll-Call Voting. American Journal of Political Science 44: Volden, Craig and Elizabeth Bergman How Strong Should Our Party Be? Party Member Preferences Over Party Cohesion. Legislative Studies Quarterly XXXI: Zeller, Belle American State Legislatures. NewYork: Crowell. 21